Mega-Evolutionary Dynamics of the Adaptive Radiation of Birds
This is a repository copy of Mega-evolutionary dynamics of the adaptive radiation of birds.
White Rose Research Online URL for this paper: http://eprints.whiterose.ac.uk/110927/
Version: Accepted Version
Article: Cooney, C.R., Bright, J.A., Capp, E.J.R. et al. (6 more authors) (2017) Mega-evolutionary dynamics of the adaptive radiation of birds. Nature. ISSN 0028-0836 https://doi.org/10.1038/nature21074
Reuse Unless indicated otherwise, fulltext items are protected by copyright with all rights reserved. The copyright exception in section 29 of the Copyright, Designs and Patents Act 1988 allows the making of a single copy solely for the purpose of non-commercial research or private study within the limits of fair dealing. The publisher or other rights-holder may allow further reproduction and re-use of this version - refer to the White Rose Research Online record for this item. Where records identify the publisher as the copyright holder, users can verify any specific terms of use on the publisher’s website.
Takedown If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing [email protected] including the URL of the record and the reason for the withdrawal request.
[email protected] https://eprints.whiterose.ac.uk/ 1 Mega-evolutionary dynamics of the adaptive radiation of birds 2 3 4 Christopher R. Cooney1§, Jen A. Bright1,2,3§, Elliot J. R. Capp1, Angela M. Chira1, Emma 5 C. Hughes1, Christopher J. A. Moody1, Lara O. Nouri1, Zoë K. Varley1, Gavin H. 6 Thomas1,4,*. 7 8 9 1Department of Animal and Plant Sciences, University of Sheffield, Sheffield, S10 2TN, 10 UK 11 2School of Geosciences, University of South Florida, Tampa, FL 33620, USA 12 3Center for Virtualization and Applied Spatial Technologies, University of South Florida, 13 Tampa, FL 33620, USA 14 4Bird Group, Department of Life Sciences, The Natural History Museum, Tring, 15 Hertfordshire, UK 16 17 18 § these authors contributed equally 19 * author for correspondence: [email protected] 20 21 22 23 24 25
1 26 The origin and expansion of biological diversity is regulated by both 27 developmental trajectories1,2 and limits on available ecological niches3-7. As 28 lineages diversify an early, often rapid, phase of species and trait proliferation 29 gives way to evolutionary slowdowns as new species pack into ever more 30 densely occupied regions of ecological niche space6,8. Small clades such as 31 Darwin’s finches demonstrate that natural selection is the driving force of 32 adaptive radiations, but how microevolutionary processes scale up to shape the 33 expansion of phenotypic diversity over much longer evolutionary timescales is 34 unclear9. Here we address this problem on a global scale by analysing a novel 35 crowd-sourced dataset of 3D-scanned bill morphology from >2000 species. We 36 find that bill diversity expanded early in extant avian evolutionary history before 37 transitioning to a phase dominated by morphospace packing. However, this early 38 phenotypic diversification is decoupled from temporal variation in evolutionary 39 rate: rates of bill evolution vary among lineages but are comparatively stable 40 through time. We find that rare but major discontinuities in phenotype emerge 41 from rapid increases in rate along single branches, sometimes leading to 42 depauperate clades with unusual bill morphologies. Despite these jumps 43 between groups, the major axes of within-group bill shape evolution are 44 remarkably consistent across birds. We reveal that macroevolutionary processes 45 underlying global-scale adaptive radiations support Darwinian9 and Simpsonian4 46 ideas of microevolution within adaptive zones and accelerated evolution between 47 distinct adaptive peaks. 48 49 The role of adaptive radiations as the source of much of the world’s biological diversity 50 has been widely emphasised10,11. Studies of small clades have provided insights into 51 the role of natural selection as a diversifying force, but cannot illuminate the processes 52 that shape the diversity and discontinuities of radiations over much longer evolutionary 53 timeframes. Indeed, at large taxonomic scales, the diversification of clades11,12 and 54 traits13 shows no evidence of the predicted slowdowns in evolutionary rates, despite 55 there being numerous examples in small clades3,14-16. This apparent paradox is 56 potentially resolved by G. G. Simpson’s model, in which major jumps to new adaptive 57 zones (“quantum evolution”) can occur unpredictably throughout clade history. These 58 jumps give rise to rapid lineage expansion into previously unoccupied niche space as 59 sub-clades continue to radiate within distinct adaptive zones and subzones4. Simpson’s 60 models introduced the concept of ‘mega-evolution’—diversification over large temporal 61 and spatial scales—unifying microevolution with other factors such as ecological 62 opportunity and evolutionary constraints that shape the macroevolutionary trajectories 63 of radiating lineages. However, while phylogenetic studies involving thousands of 64 species have demonstrated heterogeneity in rates of phenotypic evolution13,17, it is 65 unclear whether the processes outlined by Simpson play an important role in large- 66 scale adaptive radiations. This is because previous studies have been unable to 67 specifically assess the macroevolutionary dynamics of ecologically relevant traits. Here 68 we study the evolution of an important ecological trait (bill shape) across an entire Class
2 69 of organisms (birds) to elucidate the processes shaping the accumulation of phenotypic 70 diversity within a global-scale adaptive radiation. 71 72 Our approach is based around a novel data set describing avian bill shape. The avian 73 bill is closely associated with species’ dietary and foraging niches16,18,19 and represents 74 a highly-adaptable ecological trait known to play a key role in classic avian adaptive 75 radiations16,18,20. We took 3D scans of museum study skins comprising >2000 species 76 (>97% of extant genera) representing the full range of bill shape diversity. We 77 landmarked bills (Extended Data Fig. 1) using a bespoke crowd-sourcing website, 78 www.markmybird.org, and quantified the bill shape morphospace of extant birds using 79 Procrustes superimposition and Principal Components Analyses (PCA, see Methods). 80 The first eight PC axes explain >99% of the total variation in bill shape (Fig. 1). PC1 81 (58% of overall shape variation) describes the volumetric aspect ratio from elongated 82 (e.g. sword-billed hummingbird, Ensifera ensifera) to stout bills (e.g. large ground finch, 83 Geospiza magnirostris) and captures the range of shape variation encompassed by 84 standard linear measurements (length, width and depth). Variation in these bill 85 dimensions may relate to fine scale division of the dietary or foraging niche among 86 closely related species, but cannot explain the diversity of shapes observed among 87 extant birds. More complex aspects of shape (42% of total variation) are explained by 88 the remaining PCs (Fig. 1), which retain high phylogenetic signal (Extended Data Table 89 1). Importantly, although these higher shape axes explain a low proportion of shape 90 variance, they capture large differences in ecologically relevant aspects of bill shape. 91 The narrow (long tail) distributions of higher shape axes, compared to the broad 92 distribution of PC1 (Extended Data Fig. 2, Extended Data Table 1), suggest that the 93 majority of species have relatively simple bill shapes and diversify in densely packed 94 regions of bill morphospace. 95 96 We tested an important prediction of Simpson’s model by evaluating how niche 97 expansion and niche packing have contributed to the accumulation of bill shape 98 disparity throughout avian evolutionary history. We estimated multivariate disparity 99 through time using ancestral state estimates derived from rate heterogeneous models of 100 trait evolution (see Methods)13. In 1 million year time slices, we calculated disparity as 101 the sum of the variances21 from the first eight shape axes. We compared observed 102 disparity through time with two null models—constant-rate (Brownian motion) and rate 103 heterogeneous trait evolution—that are unbiased with respect to niche filling processes 104 (see Methods). Relative to these null expectations, we find that the filling of avian bill 105 morphospace through time shows a striking dominance of niche expansion early in 106 avian history, followed by a more recent transition towards niche packing (Fig. 2a-b, 107 Extended Data Fig. 2). Our data includes only extant taxa due to the poor preservation 108 of bills in the avian fossil record22, although we acknowledge that some extinct taxa had 109 bills that may lie outside the range of extant diversity (e.g. Phorusrhacidae, 110 Gastornithidae, Dromornithidae). This can result in underestimates of disparity 111 particularly if these morphologies arise early in clade history22-24. Our analyses are 112 therefore conservative with respect to transitions from bill morphospace expansion to
3 113 filling and consistent with recent studies of avian skeletal material22. The transition in the 114 mode of niche filling is consistent with a process of ever-finer divisions of niche space 115 and would be expected to correspond to slowdowns in rates of bill evolution. However, 116 the switch from niche expansion to niche packing does not map onto temporal trends in 117 the rate of bill shape evolution. Plotting evolutionary rates through time reveals an initial 118 low rate followed by a moderate (two to four-fold) increase that is coincident with the 119 divergence of many non-Passerine orders (Fig. 2c, Extended Data Fig. 3, 4). Thereafter 120 average rates dip and then rise gradually with less than 1.5-fold total variation over ~80 121 million years of evolutionary history, contrasting sharply with >250-fold variation in 122 evolutionary rate among individual lineages (Fig. 3). 123 124 The disjunction between rates of evolution and the accumulation of bill shape disparity 125 suggests that temporal trends in evolutionary rate are not necessarily indicative of the 126 underlying mode of niche filling. This decoupling could arise if some clades diverge 127 rapidly within regions of morphospace that are occupied by other clades, but where the 128 respective clades occur in allopatry. To test this idea, we mapped rates of bill evolution 129 onto the avian phylogeny (Fig. 3, Extended Data Fig. 3-5). We find several instances of 130 clades exhibiting exceptionally high rates of evolution consistent with speciational or 131 phyletic evolution within adaptive subzones (Fig. 3). Some of the fastest rates of bill 132 evolution arise in island radiations of passerine birds, where ecological divergence has 133 been closely linked to ecological opportunity (e.g. Malagasy vangas16, Galapagos 134 finches18, Hawaiian honeycreepers20), suggesting that lineages radiating on isolated 135 island archipelagos can explore morphological space independently of the global 136 avifauna. Notably high rates of bill evolution occur in several large species-rich clades 137 that have high speciation rates, including the Psittaciformes, the Furnariidae, and the 138 Passeroidea. However, these clades occupy regions of morphospace that overlap with 139 other more slowly evolving clades and so, while rapid divergence among close relatives 140 within a subzone leads to locally high rates, they do not contribute uniquely to the global 141 expansion of morphospace. In contrast, some large (Anseriformes) and some smaller 142 clades (Alcidae, Bucerotiformes) that exploit more unusual ecological resources have 143 also evolved rapidly. 144 145 Next, we find evidence for several notable instances of exceptionally high rates of 146 evolution along single branches (Extended Data Table 2). Such instances indicate 147 unusually large jumps in bill phenotype and many of the most extreme shifts (e.g. 148 Phoenicopteridae, Musophagidae, Pelecanidae, and Caprimulgiformes; Fig. 3) occur 149 towards the base of the avian radiation, consistent with the idea of early, rapid quantum 150 evolution into new adaptive zones. In some cases (e.g. Pelecanidae and Ciconiidae), 151 the evolution of extreme bill shapes is associated with a subsequent slowdown in the 152 rate of bill shape evolution (Fig. 3), suggesting that ancestral shifts towards a highly 153 specialised bill phenotype may often constrain further opportunities for either bill 154 evolution or speciation25. In contrast, some rapid jumps result in speciose clades 155 occupying more densely packed regions of morphospace. For instance, the 156 Hirundinidae diverge from other Sylvoidea but converge on a swift-like aerial insect
4 157 hawking form. These latter types of shift do not appear to be restricted to any particular 158 time periods or regions of the avian phylogeny. Similarly, the Trochiliformes diverge 159 rapidly away from the Apodiformes towards a range of bill phenotypes that opened up 160 additional opportunities for continued diversification, consistent with the idea of rapid 161 speciation driven by ecological opportunity following the invasion of an unoccupied 162 adaptive zone4,8. 163 164 Major phenotypic shifts early in the avian adaptive radiation followed by limited 165 divergence within sub-clades, implies a disconnect between mega-evolutionary 166 radiations on a global scale and adaptive radiations within smaller constituent clades. 167 Although the average phenotypes (morphospace centroids) of some higher taxa diverge 168 from one another (Extended Data Fig. 6, 7), it is unclear whether the primary axes of bill 169 shape variation within sub-clades parallel the major axes of variation across birds as a 170 whole (i.e. higher PCs), or whether evolution within clades occurs along axes of 171 variation that are distinct from the major global axes (i.e. lower PCs). We explored these 172 ideas by quantifying the variances and covariances (termed P matrices, see Methods) 173 of bill shape axes within higher taxa (families, superfamilies and orders)26,27. We find 174 that shape variation within higher taxa is explained by a single significant eigenvector of 175 P, with the exception of the Psittaciformes (two significant eigenvectors). In contrast, the 176 number of significant eigenvectors across all birds combined is three, suggesting that 177 there is low dimensional divergence within clades but high dimensional divergence 178 between clades. We then asked whether the dominant eigenvector within each sub-
179 clade (Pmax) was consistent across higher taxa. We find that bill shape (i.e. PC) axes 1 180 and 2—those that explain the majority of variation across birds as a whole—also
181 consistently load most heavily onto Pmax within higher taxa (Extended Data Fig. 7). This 182 suggests that bill shape evolution within higher taxa tends to fall back to limited 183 pathways irrespective of the position of the clade in morphospace 184 185 The low dimensionality and consistency of bill shape variation within clades, and high 186 dimensionality among clades, demonstrates striking discontinuities between how 187 phenotypic disparity accumulates in the early stages of major radiations, versus how 188 disparity accumulates as younger clades evolve within an already mature and 189 ecologically diverse radiation. This early expansion of morphospace has parallels with 190 observations of peak disparity early in clade history in palaeontological datasets of a 191 wide range of metazoan taxa28. The earliest known fossil assemblages of the ancestors 192 of modern birds, dating from the Early Cretaceous, were functionally and ecologically 193 depauperate29. It is likely that the rise of modern birds from the late Cretaceous onwards 194 occurred in a rapidly changing world30, coinciding with extensive ecological opportunity. 195 Our results imply that this dynamic adaptive landscape may have driven Simpsonian 196 mega-evolution across adaptive zones, later giving way to smaller scale fine-tuning of 197 the bill as avian diversity expanded across the globe. 198 199
5 200 Methods 201 202 Data sampling. We measured 2,028 species, representing 2,028 of 2,091 genera 203 across 194 families. Specimens were obtained primarily from the avian skin collection at 204 the Natural History Museum, Tring, and also from the Manchester Museum. Study 205 skins, rather than skeletal material, were used because they are generally much better 206 represented in museum collections with more species and specimens available than in 207 skeletons, and secondly because the rhamphotheca (the keratinous sheath surrounding 208 the fused premaxilla, maxilla and nasal bones) is often absent from skeletonised 209 specimens. This is the portion of the bill that interacts directly with the environment and 210 is thus the subject of selection. Where available, one mature male per species was 211 selected for scanning. This was necessary to achieve the taxonomic sampling required 212 within a reasonable time frame and because males are generally better represented in 213 the collections than females. Care was taken to select specimens that were 214 undamaged, with all the landmarks visible and unobstructed (see below). When 215 undamaged males were unavailable, females were preferentially chosen over unsexed 216 specimens. Some species (e.g. Strigiformes, Podargidae, and others) have bills that are 217 obscured by protruding feathers or rictal bristles that ‘shade’ the bill from the scanner. 218 For specimens where this was an issue, or for specimens that were not represented in 219 the skins collections, specimens were chosen from the skeletons collection at Tring. 220 221 3D scanning and processing. 3D scans of the bills were taken using white or blue 222 structured light scanning (FlexScan3D, LMI Technologies, Vancouver, Canada). The 223 use of 3D scans provides a more complete and nuanced estimate of bill diversity than 224 standard linear measures (length, width, depth) that reflect only the relative proportions 225 of the bill and effectively assume that bills are no more than proportional variations on a 226 cone shape. For bills of lengths > 5 cm, a R3X white-light scanner (calibration boards 10 227 – 25 mm, resolution 0.075 mm) was used, and for bills of lengths < 3 cm a MechScan 228 white-light macro scanner (calibration boards 1.3 – 4 mm, resolution 0.010 mm) was 229 used. For bills intermediate between these lengths, a pre-calibrated HDI blue-light 230 scanner (resolution 0.080 mm) was used. In some cases, larger bills (e.g. those with a 231 high aspect ratio, such as hummingbirds) were scanned on the higher resolution 232 scanner. In order to fully capture 3D geometry, approximately 5 - 25 scans per bill were 233 obtained, and aligned and combined in the FlexScan software before being exported as 234 .ply files. Scans were imported into Geomagic Studio (3D Systems, Rock Hill, SC, 235 USA), automatically decimated to approximately 500,000 faces, and cleaned to remove 236 mesh errors (holes, reversed normals, high aspect ratio spikes). In some specimens, it 237 was necessary to remove feathers or scanning artefacts that had obstructed portions of 238 the geometry by manual cleaning of the mesh. Following cleaning, meshes were 239 exported as .obj files. 240 241 Landmark choice. Landmark-based geometric morphometrics (GM) is a method for 242 analysing variation in geometric shape based on the positions of equivalent homologous 243 points (landmarks) placed on every specimen in the study31,32. While homologous in this
6 244 context is usually taken to mean developmentally homologous, in practice the key to 245 landmark selection is that the points chosen must be easily identifiable, such that they 246 can be accurately placed and repeatable within and between specimens32. This is 247 difficult to do on the rhamphotheca because, other than the tip of the bill, it lacks any 248 obvious landmarks, especially as the nostrils are not exposed in many bird species. We 249 therefore opted to identify four true landmarks: 1) the tip of the beak; and the posterior 250 margin of the keratinous rhamphotheca, along the 2) midline dorsal profile; 3) left; and 251 4) right tomial edges. Three semilandmark curves joined point 1 to points 2, 3, and 4 to 252 represent the dorsal profile, and the left and right tomial edges respectively (Extended 253 Data Fig. 1). 254 255 Crowdsourcing. In order to facilitate landmarking of such a high number of species, a 256 crowdsourcing website, www.markmybird.org, was developed to allow members of the 257 public to participate in the research by placing landmarks on to the bill scans. After 258 registration, volunteers were required to landmark two training bills with easily 259 identifiable (shoebill, Balaeniceps rex) and more challenging (brown-chested alethe, 260 Alethe poliocephala) landmarks. Instructions were shown to all users for every 261 landmark, with links to more detailed instructions provided. Bills were assigned to users 262 by randomly selecting a bill from the 100 scans most recently uploaded. To account for 263 the fact that different users will always place homologous landmarks in slightly different 264 places33, each bill was marked by three to four different users. 265 266 Quality control and landmark averaging. Custom R scripts were used to check for 267 common mistakes that may not have been caught by real-time error checks (confusing 268 left and right, large asymmetries in landmark position, incorrect order of semilandmarks, 269 and semilandmarks that deviated from the correct curve due to user failure to rotate the 270 bill and assess their landmark placement in three dimensions). If any landmark 271 configuration failed these tests, the data was manually checked and if necessary 272 removed with the bill made re-available for landmarking. Finally, the three/four 273 repetitions for each bill were averaged to find the mean shape between users, and 274 tested to ensure that all users had placed the landmarks within an acceptable range 275 (Procrustes distance < 0.2) of one another. The average bill shapes were then passed 276 forward for geometric morphometric (GM) analysis. Using ANOVA approaches for 277 assessing measurement error in geometric morphometrics33, we found that repeatability 278 was consistently high among users when comparing among PC axes (see below; 279 Extended Data Table 2). 280 281 Geometric morphometrics. All GM analysis was performed in the R package 282 Geomorph34. First, landmark configurations were subjected to a Generalised Procrustes 283 Analysis (GPA) to remove the effects of size and translational and rotational position on 284 the landmark configurations. This is a common first step in GM analyses as it removes 285 all the geometric information from the landmark coordinates that is not related to 286 shape31. During alignment, symmetry was enforced so that slight user-introduced 287 differences in the left/right positions of landmarks were removed. Semilandmarks were
7 288 slid to minimise bending energy35. The Procrustes aligned coordinates were then 289 assessed using PCA to identify the major axes of shape variation within bird bills, which 290 were plotted as morphospaces. PC scores for the first eight axes are available as 291 supplementary material. As morphospaces are projections of multidimensional Kendall’s 292 shape space into two-dimensional tangent space, they may be prone to distortions the 293 further one moves from the central coordinates of the morphospace. In other words, 294 extreme bill morphologies plotting at the edges of morphospace have the potential to 295 distort the projection such that Procrustes distances at the edges of a morphospace are 296 not equivalent to those at the centre of a morphospace. To assess the extent to which 297 projected tangent space differed from the underlying Kendall’s shape space, the 298 Procrustes aligned coordinates were analysed using tpsSmall 1.3036. We found no 299 evidence of distortion: distance in tangent was very tightly correlated with Procrustes 300 distance (uncentred correlation: 0.999; regression through the origin slope: 0.985; root 301 mean squared error < 0.001). Similarly, Procrustes distances were consistently close to 302 tangent distances (minimum Procustes D: 0.024, minimum Tangent d: 0.024; mean 303 Procustes D: 0.194, mean Tangent d: 0.192; maximum Procustes D: 0.525, maximum 304 Tangent d: 0.501). 305 306 Warps of the associated shape changes with each PC were generated by transforming 307 the landmarks of the bill closest to the average shape (rusty-fronted barwing, Actinodura 308 egertoni) to landmarks representing the extremes of a given PC when all other PCs = 0, 309 and interpolating the surface in between. 310 311 To assess any possible distortion of PCA by the underlying phylogenetic non- 312 independence among species, we also ran a phylogenetic PCA37,38. As with the 313 standard PCA, the first eight PCs accounted for >99% of total shape variance. We 314 found that the first two pPCs did not correlate with the first two original PCs—pPC1 was 315 more closely correlated with PC2 and pPC2 was more closely correlated with PC1. The 316 remaining PCs and pPCs were closely correlated and retained the same order in terms 317 of the proportion of variance explained. We also re-ran rate variable models on the first 318 eight pPCs (see below). For this analysis we allowed the pPCs to be correlated 319 because a property of pPCA is that the axes are not expected to be orthogonal. The 320 multivariate results are similar regardless of the choice of PCA or pPCA (Extended Data 321 Fig. 3). Recently identified problems inherent with using PCA (or pPCA) that can lead to 322 misidentifying macroevolutionary models are expected to arise when individual PCs are 323 analysed, particularly when the variance explained is distributed fairly evenly across 324 multiple PCs39. Because we use a multivariate approach these problems are minimized. 325 326 Phylogenetic framework. We base our analyses on the phylogenetic tree distributions 327 from www.birdtree.org11. For both ‘Hackett’ and ‘Ericson’ backbones, we sampled 328 10,000 ‘stage 2’ trees (i.e. those containing all 9,993 species) from www.birdtree.org, 329 which were pruned to generate tree distributions for the 2,028 species in our dataset. 330 We also generated similar tree distributions using ‘stage 1’ trees from the same source, 331 which contain only the subset of species placed using genetic data. Of the 2028 species
8 332 in the full dataset, 1,627 (80%) were represented in stage 1 trees. Based on these 333 distributions, we used TreeAnnotator40 to generate maximum clade credibility (MCC) 334 trees, setting branch lengths equal to ‘Common Ancestor’ node heights. In addition, we 335 constructed a composite of the Jetz et al. trees and the genomic backbone tree of Prum 336 et al.41 (Extended Data Fig. 4) by grafting sub-clades of the Stage 2 Hackett MCC tree 337 onto nodes in the Prum et al. phylogeny at positions where the two trees could be 338 sensibly combined (see Supplementary Material for node matching data and R code to 339 combine the trees). This process resulted in a composite tree combining the genus level 340 resolution afforded by the Jetz et al. tree with the branching topology and age estimates 341 of the Prum et al. backbone, which are notably younger than those in the Jetz et al. 342 trees. 343 344 Phylogenetic signal. We calculated the phylogenetic signal of bill shape by estimating 345 Pagel’s λ using the R package MOTMOT42. λ can vary between 0 and 1, with a value of 346 0 indicating no phylogenetic signal and a value of 1 indicating similar levels of 347 phylogenetic covariance as expected under a BM model. 348 349 Models of trait evolution. Univariate variable rates models were estimated using the 350 software BayesTraits (available from http://www.evolution.rdg.ac.uk/) using default 351 priors and a single-chain Markov chain Monte Carlo (MCMC) run for at least 1 billion 352 (1,000,000,000) iterations. From each chain we sampled parameters every 100,000 353 iterations and final parameter estimates for each model were based on 5,000 post-burn 354 in samples. Uncorrelated multivariate models were estimated using the same approach. 355 At each iteration in the MCMC chain, the multivariate models fit a single branch length 356 transformation to the tree across all trait (i.e. PC) axes. An uncorrelated multivariate 357 model is justified because PC axes are inherently orthogonal, however this may limit 358 inference of some forms of rate change. Specifically, the uncorrelated multivariate 359 model is informative with respect to changes in the variances among clades and shifts 360 in the morphospace centroids of clades (i.e. single branch shifts) but cannot detect 361 cases where variances and centroids are similar but covariances among clades differ. 362 We summarised the results of each run by calculating (i) the mean rate and (ii) the 363 probability of a rate shift (branch or clade) over all posterior samples for each node in 364 the tree. It is often challenging to pinpoint the precise location of rate shifts in the tree, 365 particularly when such shifts involve clades of species with short internode intervals at 366 their base. In such cases it becomes difficult to assign the location of a shift to a single 367 node and the inference of a rate shift is then often distributed across two or more nested 368 nodes in the phylogeny. To account for this, we also summarised our results using a 369 second approach in which the posterior probability for a particular rate shift was 370 calculated as the sum of the probability of a shift having occurred on a focal node or on 371 either of the nodes immediately descending from it. We focus on the multivariate 372 analyses because bill shape is a high dimensional trait. In the main text (Fig. 2, 3) we 373 report results from the stage 2 Hackett tree but found comparable results regardless of 374 tree choice (Extended Data Fig. 3, 4). 375
9 376 We checked for biases in rate estimates across the phylogeny by comparing our 377 observed multivariate rate estimates of bill shape evolution to results generated using 378 simulated data. Using the stage 2 Hackett MCC tree, we generated 10 null multivariate 379 data sets (simulated under BM) and estimated rates using runs of 200 million iterations 380 and 1,000 post-burn samples. We found that on average branch-specific rates derived 381 from simulated data sets were uncorrelated with observed rates of bill shape evolution 382 (Spearman’s rho = 0.03; p = 0.34), indicating that our results are unlikely to be affected 383 by underlying biases in rate estimation. 384 385 In addition to BayesTraits we compared the fit of three single process models (Brownian 386 motion [BM], early burst [EB] and Ornstein-Uhlenbeck [OU]), fit using the ‘fitContinuous’ 387 function and default settings in the R package Geiger v2.043, as well as alternative 388 formulations of the BAMM model44 that differed in their handling of temporal rate 389 variation (time constant [T constant], time variable [T var] and time flip [T flip]). The 390 BayesTraits, BAMM and single process models are not fitted in common a framework 391 with consistent likelihood calculations. We therefore compared the fit of the alternative 392 models within each shape axis by calculating the likelihood of a BM model fit to the 393 mean rate-transformed Jetz et al. trees derived from each model. In the absence of 394 support for alternative models (Extended Data Table 3), and because BAMM does not 395 currently allow analyses of multivariate data, we focus our interpretation on analyses 396 using BayesTraits. 397 398 Disparity and rates through time. Estimating ancestral disparity. We estimated 399 ancestral values for each component axis of bill shape variation using a maximum 400 likelihood approach implemented in the R package phytools38. We estimated ancestral 401 states using the mean rate-transformed trees for each component axis to account for 402 unequal rates of evolution across the tree and among shape axes. To generate 403 estimates of ancestral disparity through time, we took time slices at 1 million year 404 intervals starting at the root of the tree. For each time slice we extracted ancestral state 405 estimates for each component axis for the lineages in the phylogeny existing at that 406 particular time point. We then quantified multivariate disparity in trait values by 407 calculating the sum of the variances across all 8 trait axes21. Unlike other disparity 408 metrics, the sum of the variances is expected to be independent of richness and 409 sensitive to changes in both expansion and packing of trait space, thus providing an 410 indication of the relative strength of these two patterns19. 411 412 Null models of morphospace filling. We generated two alternative null models of 413 morphospace filling based on BM models of trait evolution to assess whether the 414 observed patterns of bill shape disparity through time were distinct from unbiased 415 patterns of disparity accumulation. In the first we assumed that trait variation 416 accumulates at a constant rate (‘CR’) that is homogeneous with respect to time and also 417 to a lineage’s position in the phylogeny. In the second we relaxed these assumptions of 418 rate constancy and instead simulated traits using the mean rate-transformed trees for 419 each axis, thereby providing a null model of disparity accumulation incorporating
10 420 variable rates (‘VR’) of trait evolution. For each model we simulated 500 replicate data 421 sets and used these to calculate two sets of null disparity through time curves using 422 identical approaches to those describe above. Irrespective of whether evolutionary rates 423 are fixed to be constant or allowed to vary, an important feature of both null models is 424 that the underlying balance between morphospace expansion and packing is expected 425 to be effectively equal and constant over time. This is due to the inherently non- 426 directional nature of trait change simulated using the BM model. Consequently, any 427 deviation in the observed rate of disparity accumulation compared to the null rates 428 suggests that one process (either expansion or packing) has dominated over the other. 429 430 Summarising evolutionary rates through time. For each 1 million year time slice, we 431 calculated the mean rate of evolution across all branches present at that time point. We 432 repeated this procedure for each tree in the posterior distribution to generate a 433 distribution of average rate estimates in 1 million year intervals. 434 435 Estimation of phenotypic variance-covariance (P) matrices. We examined the 436 consistency of bill shape evolution within and among avian clades using Bayesian 437 estimates of phenotypic variance-covariance matrices (P matrices) of bill shape within 438 higher taxa (families, superfamilies and orders)26,27. First, we estimated the number of 439 independent axes (i.e. eigenvectors of P) that are required to adequately explain the 440 total trait variance in P in each higher taxon. We then tested whether the dominant
441 eigenvector of bill shape variation (Pmax) is consistent among clades. Pmax is the first 442 principal component of P and an estimate of the major axis of phenotypic variation. We 443 estimated phenotypic variance-covariance matrices for higher taxa containing ≥20 444 sampled species. Posterior distributions of variance-covariance matrices were 445 generated using Bayesian MCMC MANOVA models implemented in the R package 446 MCMCglmm27. We used weak uniform priors and ran each model for 80,000 iterations 447 with a burn-in of 40,000 and sampling that produced 1,000 estimates of the posterior 448 distribution. Based on these distributions we used a set of Bayesian matrix 449 quantification approaches26 to extract information on (i) centroid position, (ii) subspace
450 orientation, (iii) individual trait loadings onto and variance explained by Pmax, and (iv) 451 number of significant eigenvectors associated with each P. 452
11 453 Figure legends 454 455 Figure 1. Bird bill morphospace density plots. PC axes 1-8 are shown as pairwise 456 scatterplots, along with warps representing the change in bill shape (n = 2028 species) 457 along each axis in dorsal and lateral views. Each axis is labeled with the proportion of 458 variance explained and estimates of phylogenetic signal (Pagel’s λ). The colour scale 459 refers to the number of species in 20 bins with minimum and maximum richness of a, 1- 460 23 b, 1-72 c, 1-64, and d, 1-98 species, respectively. 461 462 Figure 2. Morphospace filling through time. a, Accumulation of multivariate disparity 463 through time in 1 million time slices (thick black line: observed data; thin black line: after 464 LOESS smoothing; blue lines: constant rate null model; red lines: variable rate null 465 model). b, Comparison of slopes (estimated in 5 million year windows) of the LOESS- 466 smoothed observed data and null models. Differences in slope above and below zero 467 indicate dominance of morphospace expansion versus morphospace packing 468 respectively. Shading indicates 95% confidence intervals. c, Mean relative rates of 469 evolution with 95% confidence intervals (grey) through time. 470 471 Figure 3. Multivariate rates of bill shape evolution. The avian phylogeny (n = 2028 472 species) coloured by estimates of the mean relative multivariate rate of bill shape 473 evolution. Grey triangles show the stem branch of clades with support for whole clade 474 shifts in evolutionary rate. Coloured circles show rate shifts on individual internal 475 branches (colour indicates the rate estimate). The relative size of triangles and circles 476 indicates the posterior probability (PP) of a rate shift. Triangles distinguish shifts on the 477 focal node (filled) and shifts at the focal node or on one of its two daughter nodes 478 (open). 479 480 Extended Data Figure 1. Positions of landmarks and semilandmarks. The image 481 shows a 3D scan of a shoebill (Balaeniceps rex) bill marked up with four fixed 482 landmarks (numbered red points) and three semi-landmark curves along the dorsal 483 profile (from points 1 to 2) and tomial edges (left from point 1 to 3 and right from point 1 484 to 4). Each curve consists of 25 semi-landmarks (black points). 485 486 Extended Data Figure 2. Morphospace density through time. Plots show the filling 487 of avian bill morphospace through time (n = 2028 species) for PCs a, 1; b, 2; c, 3; d, 4; 488 e, 5; f, 6; g, 7; and h, 8. Densities were calculated in 1 million year time slices based on 489 univariate rate heterogeneous models of trait evolution using a stage 2 Hackett MCC 490 tree from www.birdtree.org. The scale runs from low density (blue) to high density (red), 491 indicating the extent of niche packing through time in different regions of bill 492 morphospace. For each axis the frequency distribution of PC scores among species is 493 also shown (grey bars). 494 495 Extended Data Figure 3. Comparison of multivariate rates of bill shape evolution 496 and disparity through time for alternative datasets. The plot shows estimates of the
12 497 mean relative multivariate rate of bill shape evolution for four alternative versions of the 498 avian phylogeny and also when using phylogenetic Principal Components (pPCs) (see 499 Methods). Shown below are plots comparing estimates of disparity and rates through 500 time derived from each dataset. For stage 2 trees n = 2028 species and for stage 1 501 trees n = 1627 species. 502 503 Extended Data Figure 4. Multivariate rates of bill shape evolution for a composite 504 tree based on the Prum et al. backbone. The avian phylogeny coloured according to 505 estimates of the mean relative multivariate rate of bill shape evolution. Grey triangles 506 show the stem branch of clades with support for whole clade shifts in evolutionary rate. 507 Coloured circles show rate shifts on individual internal branches (colour indicates the 508 rate estimate). The relative size of triangles and circles indicates the posterior 509 probability (PP) of a rate shift. Filled and open triangles distinguish between shifts on 510 the focal node (filled) and shifts that occur either at the focal node or on one of the two 511 immediate daughter nodes (open). 512 513 Extended Data Figure 5. Phylogenetic mapping of univariate rates of bill shape 514 evolution. The plots shows the avian phylogeny of all taxa included in the study (n = 515 2028 species) with branches coloured on a common scale across panels according to 516 estimates of the univariate rate of bill shape evolution. a, PC1, b, PC2, c, PC3, d, PC4, 517 e, PC5, f, PC6, g, PC7, h, PC8. 518 519 Extended Data Figure 6. Morphospaces of avian higher taxa. Pairwise scatterplots 520 of PCs 1 and 2, 3 and 4, 5 and 6, and 7 and 8 showing focal higher taxa (non- 521 passerines, purple; passerines, green) against total avian morphospace (grey). Values 522 in parentheses show the number of species sampled. 523 524 Extended Data Figure 7. Morphological subspaces of the P of avian higher taxa. 525 The figure shows representations of P for avian higher taxa with ≥20 species sampled. 526 First column: distribution of species values on each of the first eight raw PCs showing 527 variation in morphospace centroid for each higher taxon. Second column: two- 528 dimensional subspace for each taxon with non-passerine (purple) and passerine (green)
529 subspaces. The x- and y-axes follow the global leading (Pmax) and secondary 530 eigenvectors. Third column: percentage of total variance explained and individual PC
531 loadings onto each taxon specific Pmax. Inset: three-dimensional subspace for all non- 532 passerines (purple) and passerines (green). Values in parentheses show the number of 533 species sampled. 534 535 536 Extended Data Table 1. Variance, repeatability and phylogenetic signal of PC 537 axes. The table shows individual and cumulative variance values, kurtosis values,
538 scores of among user repeatability (R) and repeatability after averaging (Rn), and 539 maximum likelihood estimates and 95% confidence intervals of Pagel’s λ for the first
13 540 eight PC’s of bill shape. λ was estimated using two different tree topologies based on 541 the Hackett and Ericson backbone trees taken from www.birdtree.org. 542 543 Extended Data Table 2. Summary of major single-lineage bill evolutionary rate 544 shifts. Table shows fold-change rate of evolution and posterior probability (PP) for 545 major (PP > 0.7 and fold-increase > 10) ancestral single-lineage shifts in rate of bill 546 shape evolution. 547 548 Extended Data Table 3. Comparison of trait models. The table shows delta likelihood 549 values for alternative models of trait evolution within each shape axis and for different 550 tree topologies. Values were generated by calculating the likelihoods of a BM model fit 551 to the mean rate-transformed trees derived from each model. 552
14 553 1 Bright, J. A., Marugan-Lobon, J., Cobb, S. N. & Rayfield, E. J. The shapes of bird beaks 554 are highly controlled by nondietary factors. Proc Natl Acad Sci U S A 113, 5352-5357, 555 doi:10.1073/pnas.1602683113 (2016). 556 2 Lamichhaney, S. et al. Evolution of Darwin's finches and their beaks revealed by 557 genome sequencing. Nature 518, 371-375 (2015). 558 3 Phillimore, A. B. & Price, T. D. Density-dependent cladogenesis in birds. PLoS Biol 6, 559 e71, doi:10.1371/journal.pbio.0060071 (2008). 560 4 Simpson, G. G. Tempo and mode in evolution. (Columbia University Press, 1944). 561 5 Ezard, T. H. & Purvis, A. Environmental changes define ecological limits to species 562 richness and reveal the mode of macroevolutionary competition. Ecol Lett, 563 doi:10.1111/ele.12626 (2016). 564 6 Price, T. Speciation in birds. 1st edn, (Roberts and Co., 2008). 565 7 Price, T. D. et al. Niche filling slows the diversification of Himalayan songbirds. Nature 566 509, 222-225 (2014). 567 8 Losos, J. B. & Mahler., D. L. in Evolution Since Darwin: The First 150 Years (eds M. A. 568 Bell, D. J. Futuyma, W. F. Eanes, & J. S. Levinton) 381-420 (Sinauer Associates, 2010). 569 9 Reznick, D. N. & Ricklefs, R. E. Darwin's bridge between microevolution and 570 macroevolution. Nature 457, 837-842, doi:10.1038/nature07894 (2009). 571 10 Alfaro, M. E. et al. Nine exceptional radiations plus high turnover explain species 572 diversity in jawed vertebrates. Proc Natl Acad Sci U S A 106, 13410-13414, 573 doi:10.1073/pnas.0811087106 (2009). 574 11 Jetz, W., Thomas, G. H., Joy, J. B., Hartmann, K. & Mooers, A. O. The global diversity of 575 birds in space and time. Nature 491, 444-448, doi:10.1038/nature11631 (2012). 576 12 Hedges, S. B., Marin, J., Suleski, M., Paymer, M. & Kumar, S. Tree of life reveals clock- 577 like speciation and diversification. Mol Biol Evol 32, 835-845, 578 doi:10.1093/molbev/msv037 (2015). 579 13 Venditti, C., Meade, A. & Pagel, M. Multiple routes to mammalian diversity. Nature 479, 580 393-396, doi:10.1038/nature10516 (2011). 581 14 Etienne, R. S. et al. Diversity-dependence brings molecular phylogenies closer to 582 agreement with the fossil record. Proc Biol Sci 279, 1300-1309, 583 doi:10.1098/rspb.2011.1439 (2012). 584 15 Rabosky, D. L. & Glor, R. E. Equilibrium speciation dynamics in a model adaptive 585 radiation of island lizards. P Natl Acad Sci USA 107, 22178-22183, 586 doi:10.1073/pnas.1007606107 (2010). 587 16 Jonsson, K. A. et al. Ecological and evolutionary determinants for the adaptive radiation 588 of the Madagascan vangas. Proc Natl Acad Sci U S A 109, 6620-6625, 589 doi:10.1073/pnas.1115835109 (2012). 590 17 Rabosky, D. L. et al. Rates of speciation and morphological evolution are correlated 591 across the largest vertebrate radiation. Nat Commun 4, 1958, doi:10.1038/ncomms2958 592 (2013). 593 18 Grant, P. R. Ecology and evolution of Darwin's finches. 2nd edn, (Princeton University 594 Press, 1999). 595 19 Pigot, A. L., Trisos, C. H. & Tobias, J. A. Functional traits reveal the expansion and 596 packing of ecological niche space underlying an elevational diversity gradient in 597 passerine birds. Proc Biol Sci 283, doi:10.1098/rspb.2015.2013 (2016). 598 20 Lovette, I. J., Bermingham, E. & Ricklefs, R. E. Clade-specific morphological 599 diversification and adaptive radiation in Hawaiian songbirds. Proc Biol Sci 269, 37-42, 600 doi:10.1098/rspb.2001.1789 (2002). 601 21 Kraft, N. J. B., Valencia, R. & Ackerly, D. D. Functional traits and niche-based tree 602 community assembly in an amazonian forest. Science 322, 580-582 (2008). 603 22 Mitchell, J. S. Extant-only comparative methods fail to recover the disparity preserved in 604 the bird fossil record. Evolution 69, 2414-2424 (2015). 605 23 Finarelli, J. A. & Goswami, A. Potential Pitfalls of Reconstructing Deep Time 606 Evolutionary History with Only Extant Data, a Case Study Using the Canidae 607 (Mammalia, Carnivora). Evolution 67, 3678-3685 (2013).
15 608 24 Slater, G. J. Iterative adaptive radiations of fossil canids show no evidence for diversity- 609 dependent trait evolution. P Natl Acad Sci USA 112, 4897-4902 (2015). 610 25 Ricklefs, R. E. Small clades at the periphery of passerine morphological space. Am Nat 611 165, 651-659, doi:10.1086/429676 (2005). 612 26 Robinson, M. R. & Beckerman, A. P. Quantifying multivariate plasticity: genetic variation 613 in resource acquisition drives plasticity in resource allocation to components of life 614 history. Ecol Lett 16, 281-290, doi:10.1111/ele.12047 (2013). 615 27 Hadfield, J. D. MCMC Methods for Multi-Response Generalized Linear Mixed Models: 616 The MCMCglmm R Package. J Stat Softw 33, 1-22 (2010). 617 28 Hughes, M., Gerber, S. & Wills, M. A. Clades reach highest morphological disparity early 618 in their evolution. Proc Natl Acad Sci U S A 110, 13875-13879, 619 doi:10.1073/pnas.1302642110 (2013). 620 29 Mitchell, J. S. & Makovicky, P. J. Low ecological disparity in Early Cretaceous birds. P 621 Roy Soc B-Biol Sci 281 (2014). 622 30 Brusatte, S. L., O'Connor, J. K. & Jarvis, E. D. The Origin and Diversification of Birds. 623 Curr Biol 25, R888-898, doi:10.1016/j.cub.2015.08.003 (2015). 624 31 Klingenberg, C. P. Visualizations in geometric morphometrics: how to read and how to 625 make graphs showing shape changes. Hystrix 24, 15-24 (2013). 626 32 Zelditch, M. Geometric morphometrics for biologists : a primer. (Elsevier Academic 627 Press, 2004). 628 33 Fruciano, C. Measurement error in geometric morphometrics. Dev Genes Evol 226, 139- 629 158 (2016). 630 34 Adams, D. C. & Otarola-Castillo, E. geomorph: an r package for the collection and 631 analysis of geometric morphometric shape data. Methods Ecol Evol 4, 393-399 (2013). 632 35 Gunz , P., Mitteroecker, P. & Bookstein, F. L. in Modern Moprhometrics in Physical 633 Anthropology (ed D. E. Slide) 73-98 (Kluwer Academic/Plenum Publishers, 2004). 634 36 tpsSmall, testing amount of shape variation, version 1.30 (Department of Ecology and 635 Evolution, State University of New York at Stony Brook, 2014). 636 37 Revell, L. J. Size-Correction and Principal Components for Interspecific Comparative 637 Studies. Evolution 63, 3258-3268 (2009). 638 38 Revell, L. J. phytools: an R package for phylogenetic comparative biology (and other 639 things). Methods Ecol Evol 3, 217-223 (2012). 640 39 Uyeda, J. C., Caetano, D. S. & Pennell, M. W. Comparative Analysis of Principal 641 Components Can be Misleading. Systematic Biology 64, 677-689 (2015). 642 40 Drummond, A. J., Suchard, M. A., Xie, D. & Rambaut, A. Bayesian Phylogenetics with 643 BEAUti and the BEAST 1.7. Molecular Biology and Evolution 29, 1969-1973 (2012). 644 41 Prum, R. O. et al. A comprehensive phylogeny of birds (Aves) using targeted next- 645 generation DNA sequencing. Nature 526, 569-573, doi:10.1038/nature15697 (2015). 646 42 Thomas, G. H. & Freckleton, R. P. MOTMOT: models of trait macroevolution on trees. 647 Methods Ecol Evol 3, 145-151, doi:10.1111/j.2041-210X.2011.00132.x (2012). 648 43 Pennell, M. W. et al. geiger v2.0: an expanded suite of methods for fitting 649 macroevolutionary models to phylogenetic trees. Bioinformatics 30, 2216-2218, 650 doi:10.1093/bioinformatics/btu181 (2014). 651 44 Rabosky, D. L. Automatic Detection of Key Innovations, Rate Shifts, and Diversity- 652 Dependence on Phylogenetic Trees. Plos One 9 (2014). 653 654 655 656 657 658
16 659 Acknowledgements 660 We thank Mark Adams, Hein van Grouw, and Robert Prys-Jones from the Bird Group at 661 the NHM, Tring and Henry McGhie at the Manchester Museum for providing access to 662 and expertise in the collections; Shai Meiri of Tel Aviv University for providing a sample 663 of study skins; Simon Stone of MechInovation Ltd. for providing training and advice on 664 3D scanning; Matthew Groves, Jamie McLaughlin, Mike Pidd of HRI Digital for the 665 construction of www.markmybird.org; Andrew Beckerman for advice on analysing P 666 matrices; Emily Rayfield, Alex Pigot, Arne Mooers, and Alex White for providing 667 valuable comments on pre-submission drafts of the manuscript. Finally, we are indebted 668 to the wonderful volunteer citizen scientists at www.markmybird.org for generously 669 giving up their time to help build the database of bird bill shape and contribute to our 670 understanding of avian evolution. This work was funded by the European Research 671 Council (grant number 615709 Project ‘ToLERates’) and by a Royal Society University 672 Research Fellowship to GHT (UF120016). 673 674 Author information 675 These authors contributed equally to this work: Christopher R. Cooney & Jen A. Bright. 676 677 Affiliations 678 Department of Animal and Plant Sciences, University of Sheffield, Sheffield, S10 2TN, 679 UK 680 Christopher R. Cooney, Jen A. Bright, Elliot J. R. Capp, Angela M. Chira, Emma C. 681 Hughes, Christopher Moody, Lara O. Nouri, Zoë K. Varley, Gavin H. Thomas 682 683 School of Geosciences, University of South Florida, Tampa, FL 33620, USA 684 Jen A. Bright 685 686 Center for Virtualization and Applied Spatial Technologies, University of South Florida, 687 Tampa, FL 33620, USA 688 Jen A. Bright 689 690 Bird Group, Department of Life Sciences, The Natural History Museum, Tring, 691 Hertfordshire, UK 692 Gavin H. Thomas 693 694 Contributions 695 Christopher R. Cooney, Jen A. Bright, and Gavin H. Thomas conceived of the study, 696 designed analytical protocols, analysed the data and wrote the manuscript. All authors 697 collected and processed data and provided editorial input into the manuscript. 698 699 Competing financial interests. None. 700 701 Corresponding author 702 Correspondence to: Gavin H. Thomas 703 704 Supplementary Information
17 705 Excel file (PC_scores_all_genera.csv): this file contains scores for the first eight non- 706 phylogenetic PC’s for all species (n = 2028) in our data. 707 708 Excel file (PrumMerge_CRC.xlsx): this file details the mapping of Jetz et al. clades to 709 the Prum et al. backbone phylogeny. The table shows the nodes used to attach patch 710 clades from the Jetz et al. stage 2 Hackett tree to the Prum et al. backbone phylogeny. 711 712 PrumMerge.zip: this archive contains data files and an R script to combine the 713 backbone (approximately family level) phylogeny of Prum et al. with the species level 714 resolution of the Jetz et al avian phylogeny. 715 716 AvianPhylogenies.zip: this archive contains all alternative genus level phylogenies used 717 in our analyses. 718 719 Raw scan data in obj format and text files containing individual and species-averaged 720 landmarks are available from the Natural History Museum Data Portal here: 721 http://dx.doi.org/10.5519/0005413. 722
18 a b 0.4 0.2 0.2 0.76)
0.88)
=
= 0.1
λ
λ
0.0 0.0 (3%,
(29%,
0.2 − 0.1 PC4 PC2 − 0.4 − 0.2 − −0.4 −0.2 0.0 0.2 0.4 −0.2 −0.1 0.0 0.1 0.2 PC1 (58%, λ = 0.95) # species PC3 (6%, λ = 0.85) 10 11 12 13 14 15 16 18 19 20 21 22 23 1 2 3 4 5 6 8 9
c d 0.10 0.05 0.81)
0.82)
0.05 =
=
λ
λ
0.00 0.00 (1%,
(<1%,
PC6 0.05 PC8 − 0.05 − 0.10 − −0.10−0.05 0.00 0.05 0.10 0.15 −0.10 −0.05 0.00 0.05 PC5 (2%, λ = 0.89) PC7 (1%, λ = 0.80)
Figure 1. Bird bill morphospace density plots. PC axes 1-8 are shown as pairwise sca3erplots, along with warps represen8ng the change in bill shape (n = 2028 species) along each axis in dorsal and lateral views. Each axis is labeled with the propor8on of variance explained and es8mates of phylogene8c signal (Pagel’s λ). The colour scale refers to the number of species in 20 bins with minimum and maximum richness of a, 1-23 b, 1-72 c, 1-64, and d, 1-98 species, respec8vely. 1 a b c null) 2 − 0.01 0.8 0 scaled) 0.6 − 1 0.01 0.4 − Disparity (re 0.2 0.02 Observed of evolution rate Relative − Null (BM) Null (VR) 0 0 Difference in slope (observed slope in Difference
100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 Time (MYA) Time (MYA) Time (MYA)
Figure 2. Morphospace filling through Picathartidae Muscicapoidea Petroicidae Corvoidea Pomatostomidae Meliphagoidea Ptilonorhynchoidea Menuroidea Passeroidea Tyrannidae Cotingidae 0.1 0.2 0.5 1 2 5 10 25 Struthioniformes Relative rate Rheiformes Apterygiformes Casuariiformes Furnariidae Tinamiformes Anseriformes Thamnophilidae Galliformes Eurylaimides Phaethontiformes Opisthocomiformes Acanthisitti Phoenicopteriformes Podicipediformes Pteroclidiformes Psittaciformes Mesitornithiformes Columbiformes Otidiformes Falconiformes Cuculiformes Cariamiformes Gruiformes Piciformes Musophagiformes Gaviiformes Sphenisciformes Coraciiformes Procellariiformes Ciconiiformes Bucerotiformes Suliformes Trogoniformes Pelecaniformes Leptosomiformes Eurypygiformes Accipitriformes Caprimulgiformes Strigiformes Apodiformes Coliiformes Charadriiformes Trochiliformes Figure 3. Mul 0.2 0.2 0.0 0.0 PC1 PC2 0.2 590 − 590 590 852 −0.2 396178 396178 178 240 12036 12036 54 54 247 0.4 247 7 − 8 15 15 −0.4 1 1 1 1 0 20 40 60 80 100 0 20 40 60 80 100 t t Counts 0.2 c d 0.2 0.1 0.1 0.0 PC4 PC3 0.0 −0.1 590 590 1114 1813 396178 396178 229 −0.1 444 12036 12036 −0.2 52 68 247 247 9 10 15 15 1 1 1 −0.2 1 0 20 40 60 80 100 0 20 40 60 80 100 t t Counts 0.15 e 0.10 f 0.10 0.05 0.05 PC6 PC5 0.00 0.00 590 590 −0.05 1048 1150 396178 396178 284 −0.05 307 12036 12036 50 53 247 247 −0.10 9 9 15 15 1 0.10 1 1 − 1 0 20 40 60 80 100 0 20 40 60 80 100 t t Counts g 0.05 h 0.05 0.00 PC8 0.00 PC7 0.05 590 590 − 1876 1414 396178 396178 455 363 12036 12036 69 59 247 247 11 −0.05 10 15 15 −0.10 1 1 1 1 0 20 40 60 80 100 0 20 40 60 80 100 t t Extended Data Figure 2. Morphospace density through 0.1 0.2 0.5 1 2 5 10 25 0.1 0.2 0.5 1 2 5 10 25 Relative rate Rate Stage 1 Hackett Stage 1 Ericson Stage 2 Hackett – pPCs 0.1 0.2 0.5 1 2 5 10 25 0.1 0.2 0.5 1 2 5 10 25 0.1 0.2 0.5 1 2 5 10 25 Rate Rate Relative rate 1 5 0.8 4 scaled) 0.6 − 3 0.4 2 Disparity (re 1 0.2 Relative rate of evolution rate Relative 0 0 100 80 60 40 20 0 100 80 60 40 20 0 Time (MYA) Time (MYA) Extended Data Figure 3. Comparison of mul Picathartidae Muscicapoidea Petroicidae Corvoidea Pomatostomidae Meliphagoidea Ptilonorhynchoidea Menuroidea Passeroidea Tyrannidae Cotingidae 0.1 0.2 0.5 1 2 5 10 25 Struthioniformes Relative rate Rheiformes Apterygiformes Casuariiformes Furnariidae Tinamiformes Anseriformes Thamnophilidae Galliformes Eurylaimides Acanthisitti Caprimulgiformes Apodiformes Psittaciformes Falconiformes Trochiliformes Cariamiformes Piciformes Musophagiformes Otidiformes Coraciiformes Cuculiformes Mesitornithiformes Bucerotiformes Pteroclidiformes Trogoniformes Columbiformes Leptosomiformes Gruiformes Coliiformes Eurypygiformes Strigiformes Accipitriformes Phaethontiformes Charadriiformes Gaviiformes Sphenisciformes Opisthocomiformes Podicipediformes Phoenicopteriformes Procellariiformes Pelecaniformes Ciconiiformes Suliformes 1 b c 4 d 0.04 null) − 0.8 0.02 scaled) 0.6 − 2 0 0.4 Disparity (re 0.02 0.2 − Observed of evolution rate Relative Null (BM) Null (VR) 0 0 Difference in slope (observed slope in Difference 0.04 − 80 60 40 20 0 80 60 40 20 0 80 60 40 20 0 Time (MYA) Time (MYA) Time (MYA) Extended Data Figure 4. Mul a b −16 −14 −12 −10 −8 −6 −4 −16 −14 −12 −10 −8 −6 −4 log ( Rate ) log ( Rate ) c d −16 −14 −12 −10 −8 −6 −4 −16 −14 −12 −10 −8 −6 −4 log ( Rate ) log ( Rate ) e f −16 −14 −12 −10 −8 −6 −4 −16 −14 −12 −10 −8 −6 −4 log ( Rate ) log ( Rate ) g h −16 −14 −12 −10 −8 −6 −4 −16 −14 −12 −10 −8 −6 −4 log ( Rate ) log ( Rate ) Extended Data Figure 5. Phylogene Galliformes 63% Furnariidae 52% (73) (84) Columbiformes 63% Cotingidae 78% (42) (37) Cuculiformes 45% Tyrannidae 53% (27) (102) Gruiformes 53% Meliphagoidea 56% (38) (59) Procellariiformes 58% Corvoidea 58% (27) (136) Pelecaniformes 65% Sylvioidea 63% (37) (208) Trochiliformes 69% Muscicapoidea 73% (102) (132) Charadriiformes 64% Passeroidea 59% (87) (296) Accipitriformes 41% (70) P Coraciiformes 66% 2 (32) P Piciformes 75% max (66) P3 Psittaciformes 44% (84) Pmax Extended Data Figure 7. Morphological subspaces of the P of avian higher taxa. The figure shows representa8ons of P for avian higher taxa with ≥20 species sampled. First column: distribu8on of species values on each of the first eight raw PCs showing varia8on in morphospace centroid for each higher taxon. Second column: two-dimensional subspace for each taxon with non-passerine (purple) P and passerine (green) subspaces. The x- and y-axes follow the global leading ( max) and secondary eigenvectors. Third column: percentage of total variance explained and individual PC loadings onto P . each taxon specific max Inset: three-dimensional subspace for all non-passerines (purple) and passerines (green). Values in parentheses show the number of species sampled. Variance Cumulative PC axis Kurtosis R R Stage 2 Hackett λ Stage 2 Ericson λ (%) (%) n 1 57.8 57.8 –0.487 0.998 1.000 0.949 (0.931-0.964) 0.954 (0.936-0.968) 2 29.0 86.8 0.795 0.913 0.976 0.758 (0.704-0.806) 0.760 (0.706-0.808) 3 6.2 93.1 1.381 0.967 0.991 0.851 (0.813-0.882) 0.861 (0.824-0.892) 4 2.8 95.9 7.370 0.987 0.997 0.878 (0.845-0.906) 0.873 (0.838-0.903) 5 1.8 97.7 1.867 0.977 0.994 0.897 (0.863-0.924) 0.888 (0.851-0.917) 6 0.9 98.6 2.122 0.945 0.985 0.822 (0.774-0.863) 0.816 (0.766-0.858) 7 0.4 99.0 6.426 0.953 0.987 0.803 (0.756-0.843) 0.803 (0.756-0.843) 8 0.3 99.2 3.452 0.938 0.983 0.805 (0.752-0.848) 0.794 (0.739-0.840) Extended Data Table 1. Variance, repeatability and phylogene repeatability aSer averaging (Rn), and maximum likelihood es8mates and 95% confidence intervals of Pagel’s λ for the first eight PC’s of bill shape. λ was es8mated using two different tree topologies based on the Hacke3 and Ericson backbone trees taken from www.birdtree.org. Fold- Order Family Genera N PP increase PHOENICOPTERIFORMES Phoenicopteridae Phoeniconaias, Phoenicoparrus, Phoenicopterus 3 45.2 1.000 APODIFORMES Trochilidae Discosura, Lophornis, Sephanoides 3 38.5 0.999 PELECANIFORMES Threskiornithidae Bostrychia, Cercibis, Eudocimus, Geron8cus, 14 29.6 0.989 Lopho8bis, Mesembrinibis, Nipponia, Phimosus, Platalea, Plegadis, Pseudibis, Thauma8bis, Theris8cus, Threskiornis PASSERIFORMES Dendrocolap8dae Campylorhamphus, Drymornis, Lepidocolaptes 3 23.5 0.994 PASSERIFORMES Paradisaeidae Paro8a, Pteridophora 2 22.2 0.992 PASSERIFORMES Melanochari8dae Oedistoma, Toxorhamphus 2 21.4 0.914 PASSERIFORMES Platysteiridae Ba8s, Platysteira 2 20.1 0.990 PICIFORMES Ramphas8dae Andigena, Aulacorhynchus, Pteroglossus, 5 18.9 0.988 Ramphastos, Selenidera ANSERIFORMES Ana8dae Lophodytes, Mergellus, Mergus 3 18.4 0.974 ACCIPITRIFORMES Accipitridae Helicolestes, Rostrhamus 2 18.0 0.980 PASSERIFORMES Hirundinidae Alopochelidon, Ahcora, Cheramoeca, Delichon, 19 14.8 0.783 Eurochelidon, Haplochelidon, Hirundo, Neochelidon, No8ochelidon, Petrochelidon, Phedina, Progne, Psalidoprocne, Pseudhirundo, Pseudochelidon, Pygochelidon, Riparia, Stelgidopteryx, Tachycineta PASSERIFORMES Fringillidae Loxioides, Telespiza 2 13.0 0.842 MUSOPHAGIFORMES Musophagidae Corythaeola, Corythaixoides, Crinifer, 6 11.5 0.838 Musophaga, Ruwenzorornis, Tauraco PASSERIFORMES Timaliidae Jabouilleia, Rimator 2 11.1 0.981 Extended Data Table 2. Summary of major single-lineage bill evolu Stage 2 Ericson 1 0 71.3 166.5 302.2 623.6 618.8 623.6 2 0 82.8 172.5 286.4 575.1 483.4 575.1 3 0 51.2 164.3 338.7 583.6 529.3 583.6 4 0 65.5 157.0 283.7 875.0 824.7 875.0 5 0 59.1 172.6 310.9 625.8 577.1 625.8 6 0 50.2 128.5 261.3 710.8 636.7 710.8 7 0 58.6 159.2 297.1 805.7 720.7 805.7 8 0 69.9 154.1 333.7 831.3 728.2 831.3 Stage 1 Hacke3 1 0 56.8 134.7 227.2 479.5 473.6 479.5 2 0 59.8 149.8 243.1 483.8 398.0 483.8 3 0 26.4 135.5 271.1 493.5 439.2 493.5 4 0 40.5 128.4 237.4 714.7 675.2 714.7 5 0 52.0 136.7 278.4 478.6 439.8 478.6 6 0 22.6 95.7 219.2 579.5 517.6 579.5 7 0 26.3 135.1 238.4 670.7 586.1 670.7 8 0 29.1 103.4 232.2 675.3 570.4 675.2 Stage 1 Ericson 1 0 69.7 132.5 248.7 486.4 479.6 486.4 2 0 59.4 143.3 239.7 488.2 400.3 488.2 3 0 21.8 136.4 275.2 502.7 447.2 502.7 4 0 32.5 132.1 245.3 721.8 679.5 721.8 5 0 53.8 130.3 275.0 482.9 442.3 482.9 6 0 23.9 90.3 233.9 583.7 519.5 583.7 7 0 34.9 132.3 243.6 669.7 585.1 669.7 8 0 29.5 101.1 244.4 676.4 569.8 676.4 Extended Data Table 3. Comparison of trait models. The table shows delta likelihood values for alterna8ve models of trait evolu8on within each shape axis and for different tree topologies. Values were generated by calcula8ng the likelihoods of a BM model fit to the mean rate-transformed trees derived from each model.